Bilateral AI dialogues between great powers may shape global AI governance more than multilateral forums. Critically examine, with reference to India's position.

At their September 2026 summit, the U.S. and China launched a dialogue on AI risks and benefits under an eight-point consensus [1]. The two leading AI powers now have a direct channel on frontier AI. Whether that channel will outweigh multilateral forums in shaping global AI governance depends on more than market power.

Why bilateral tracks may dominate

• Concentration of capability: Most frontier models, advanced chips and compute sit in these two countries. What they agree between themselves can become the default rule for everyone else.
• Frontier focus: Beijing said it valued Washington's use of the term "super intelligence" instead of "artificial intelligence" [1]. This shows both sides want to talk about the highest-risk systems, which multilateral forums have found hardest to handle.
• Speed and flexibility: Two leaders can move faster than 193 states. The same summit also set up a trade council [1], which shows how quickly a bilateral track can create an institution.

Why this claim is overstated

• Dialogue is not regulation: The AI track has no stated commitments, limits or checking system [1]. Rivalry between the two also makes it fragile.
• Multilateral structures are already in place: UNGA Resolution A/RES/79/325 created the Global Dialogue on AI Governance, which first met in Geneva in July 2026, and an Independent International Scientific Panel on AI [2].
• Legitimacy lies in broad rules: In trade, more than 80% of world goods trade still runs on WTO MFN terms despite many bilateral deals [3]. Shared rules outlast deals made between two countries.

India's position

• Voice for the Global South: The New Delhi Declaration from the India AI Impact Summit (Feb 2026) was endorsed by 92 countries and organisations under the principle "Sarvajan Hitaya, Sarvajan Sukhaya" [4].
• Frontier engagement: It also produced the New Delhi Frontier AI Impact Commitments from 13 model developers and added sovereign GPU capacity [4].
• Risk: If the U.S. and China settle frontier-AI rules on compute, data and safety testing between themselves, India could end up accepting rules it did not help write.

In short, bilateral great-power dialogues can set the agenda on frontier risks, but the UN and summit-based processes still give AI rules their legitimacy. India should link the two tracks. It can push for bilateral safety understandings to be brought into the UN Global Dialogue, and use its convening role to keep AI governance inclusive and equitable, in line with SDG 10 and the Global Digital Compact.